In my experience, I always get much better Searx results by changing the language to the relevant one (and disable SafeSearch). However, between instances, it varies which search sources they include (Google, Bing, DDG etc.). After trying multiple, I’ve found a few instances that generally get me the results I’m looking for.

E.g. if searching for, let’s say, something German, my URL would contain this:

/search?q=cannabis+legal&language=de&safesearch=0

I don’t feel too inclined to argue with them

Interestingly, that makes me feel a bit inclined to argue with you about this.

To me, it just feels like…

“No, you can’t upgrade your old car speakers, because the custom ones would be too loud.”

“No, you can’t upgrade the seats, because the 3rd party ones are way too comfortable.”

“No, you can’t keep that Wunderbaum dangling under the mirror, it modifies the smell and it’s too damn fresh for you!”

Well, actually, it’s more like…

“No, silly, you simply can’t use those carefully crafted custom-made playing pieces, cards or high-DPI printed board to extend/change this board game to your liking, since we didn’t sell you any of that…”

As you see it, what’s the difference? Or would you in fact just happily accept those scenarios too?

Was that the one with too many pancakes?

You’re always bringing something interesting to look into or write on the list for later. This time too! Nimletter is on my list now.

Because it has fewer parameters and (in some cases) it’s quantized. The hardware needed to run local inference on the full model is not really feasible to most people. Though, the release of it will probably still make a wide impact on the quality of other upcoming smaller models being distilled from it, or trained on synthetic data from it, or merged with it, etc.

You’re asking the real important questions!

And it’s great for sorting by date.